Projects using OpenAI in LondonProjects using OpenAI in LondonRelay
WhatsApp on autopilot · books, quotes, qualifies, never misses a message
Most service businesses lose money in WhatsApp. The phone sits on someone's desk, replies happen between meetings, the same five questions get asked all day, and half the leads never make it to the CRM. By the time anyone gets back to a new enquiry, the customer has booked with whoever replied first.
Relay turns WhatsApp into a real channel.
The business gets a number. Customers message it. An AI agent answers in your tone, in any language, around the clock. It qualifies leads against your rules, books straight into your calendar, sends quotes from your price list, takes deposits via Stripe, and writes everything back to your CRM. A human only steps in when the conversation actually needs them.
What you get
24/7 first response on every WhatsApp enquiry, in your voice
Bookings landing directly in your calendar, not in someone's notebook
Quotes sent in seconds with your pricing rules applied
Deposits collected before the customer hangs up
Full conversation history per customer, searchable, in your CRM
Clean human handover when a conversation needs you
Who it's for
Service businesses where WhatsApp is the main inbound channel. Car rental, salons, clinics, trades, removals, holiday lets, restaurants. Anyone whose customers expect a reply faster than email allows.
What it replaces
A part-time receptionist, a £200/month chatbot, or you answering at 11pm. Same outcome, one-off setup, no per-message charge.
Production reality
Live with paying businesses today. Multi-tenant from day one, so each business gets its own tone, prices, calendar, and bank account, fully isolated. Standard deploy from kick-off to live in 7 days. Designed and built a production-ready internal AI support bot from scratch. The system uses a five-layer RAG architecture to ingest and retrieve knowledge from different knowledge sources, e.g. Confluence, Jira, Notion, and Slack, enabling staff to get accurate, source-cited answers to internal queries directly within Slack.
Built entirely on internal infrastructure using the Claude API, Supabase, and pgvector - ensuring zero data exposure outside the organisation. The knowledge base spans 10,000+ entries across 4 data sources, with hybrid retrieval combining semantic search, keyword fallback, and query rewriting for accuracy.
The bot replaced a manual support model, reduced repetitive query handling, and introduced a scalable self-service layer for internal IT support / helpdesk - while resolving the data privacy concerns that had blocked a third-party vendor solution.
